Projections of Earth’s Technosphere: Luminosity and

Mass as Limits to Growth

Jacob Haqq-Misraa,∗, Clément Vidalb , George Profitiliotisa
arXiv:2410.23420v1 [astro-ph.EP] 30 Oct 2024

a
Blue Marble Space Institute of Science, Seattle, WA, USA
b
Center Leo Apostel, Vrije Universiteit Brussel, Brussels, Belgium

Abstract
Earth remains the only known example of a planet with technology, and
future projections of Earth’s trajectory provide a basis and motivation for
approaching the search for extraterrestrial technospheres. Conventional ap-
proaches toward projecting Earth’s technosphere include applications of the
Kardashev scale, which suggest the possibility that energy-intensive civi-
lizations may expand to harness the entire energy output available to their
planet, host star, or even the entire galaxy. In this study, we argue that the
Kardashev scale is better understood as a “luminosity limit” that describes
the maximum capacity for a civilization to harvest luminous stellar energy
across a given spatial domain, and we note that thermodynamic efficiency
will always keep a luminosity-limited technosphere from actually reaching
this theoretical limit. We suggest the possibility that an advanced techno-
sphere might evolve beyond this luminosity limit to draw its energy directly
from harvesting stellar mass, and we also discuss possible trajectories that
could exist between Earth today and such hypothetical “stellivores.” We de-
velop a framework to describe trajectories for long-lived technospheres that
optimize their growth strategies between exploration and exploitation, un-
like Earth today. We note that analyses of compact accreting stars could
provide ways to test the stellivore hypothesis, and we more broadly suggest
an expansion of technosignature search strategies beyond those that reside
exactly at the luminosity limit.

∗
Corresponding author
Email address: [email protected] (Jacob Haqq-Misra)

Preprint submitted to Acta Astronautica November 1, 2024

1. Introduction
The search for technosignatures depends on making meaningful projec-
tions of Earth’s future in order to inform our thinking and planning. Projec-
tions of the future in the context of the search for technosignatures often as-
sume continuous exponential growth and expansion across the galaxy, which
is based on a model that was developed by Kardashev [1]. The original Kar-
dashev scale included a Type I civilization that has harnessed all the power
available to its planet, a Type II civilization that has harnessed all available
stellar power, and a Type III civilization that has harnessed the power of all
stars in a galaxy. The Kardashev scale remains an important reference point
for discussing possible observational characteristics and motivating actual
searches for such energy-intensive civilizations.
The idea of continuous growth that underlies the Kardashev scale con-
tains two embedded assumptions: (1) growth in energy use and (2) growth
in spatial scale. Kardashev assumed that human civilization would continue
increasing its energy consumption at a 1% annual growth rate, which would
eventually reach a point at which all sunlight falling on the planet (Type I) or
all sunlight emitted from the star (Type II) would be required. The original
intent of Kardashev [1] was to place constraints on the transmitting power
for interstellar communication, which was focused on growth in energy use
that corresponded to the different spatial scales. However, many subsequent
analyses have assumed that this projected rate in annual energy use is cor-
related with an increase in the spatial domain available to the civilization;
or in other words, a civilization with exponentially increasing energy needs
is assumed to also proportionally expand across the galaxy. These two as-
sumptions are taken together in many applications of the Kardashev scale
[see, e.g., 2, for an overview], but it is not necessarily evident that future
changes in the energy use or spatial domain of Earth (or an extraterrestrial
technosphere) must follow this trajectory.
This paper suggests that the Kardashev scale is best interpreted as a
“luminosity limit,” which describes the maximum capacity for harnessing lu-
minous stellar energy across a spatial domain. We retain the assumption that
stars are the primary source of energy available to long-lived technospheres.
We suggest that many plausible technosphere scenarios could exist below the
luminosity limit. We also discuss potential cases in which other methods of
harvesting stellar energy could exceed the luminosity limit and approach a
limit based on the availability of mass in a system.

2
 Scenario Annual Energy Use (J) Domain Radius (m) Domain Description
Present-day
Earth 6 × 1020 6.4 × 106 Earth
Extended Kardashev Scale
Type I 3 × 1023 6.4 × 106 Planet
Type II 3 × 1033 3.5 × 1013 Star
Type III 3 × 1043 2 × 1020 Galaxy
Type IV 3 × 1053 1027 Universe
1000-year Projections
Living with the Land (S4) 3 × 1019 6.4 × 106 Earth
Restoration (S7) 3 × 1020 6.4 × 106 Earth
Golden Age (S3) 8 × 1020 4.5 × 1012 Venus to Neptune
Ouroboros (S8) 2 × 1021 3.8 × 108 Earth to Moon
Big Brother is Watching (S1) 2 × 1021 4.5 × 1012 Venus to Neptune
Out of Eden (S10) 2 × 1021 2 × 1013 Mercury to Heliopause
Transhumanism (S5) 3 × 1021 7.4 × 1012 Venus to Kuiper Belt
Wild West (S2) 8 × 1021 3.4 × 1011 Earth to Asteroid Belt
Sword of Damocles (S6) 2 × 1022 7.4 × 1012 Venus to Kuiper Belt
Deus Ex Machina (S9) 1025 2 × 1013 Mercury to Heliopause

Table 1: Energy use and spatial domain for present-day Earth, the extended Kardashev
scale [3], and projections of Earth’s 1000-year future [4] sorted by ascending annual energy
use.

2. The Luminosity Limit

The basic idea underlying our approach is that a technosphere, as well as a
biosphere, needs both energy and resources. The Kardashev scale links both
of these requirements, but we present a framework to more explicitly show
the relationship between these variables. We consider several technosphere
scenarios in terms of the annual energy use (in Joules) and the spatial domain
radius (in meters). The spatial domain describes the largest physical volume
in which a technosphere can obtain resources; it is not necessarily required
that this entire volume is inhabited or fully utilized, but only that it is
utilizable. We summarize all the scenarios considered in Table 1 and Figure
1.
The values for energy use for the Kardashev scale follow the extended
classification scheme developed by Gray [3], which includes a Type IV civ-
ilization that has harnessed the power of the entire universe. The domain
radius for the Type I, II, III, and IV scenarios is the maximum spatial extent
required to achieve the corresponding annual energy use from harnessing
stellar luminosity. This enables the Kardashev scale to be expressed as a
function of domain radius and annual energy use, which is shown in Figure
1 as the “luminosity limit.” We can fit a line to the luminosity limit to find
an approximate functional relationship between the maximum annual energy

3
 Planet Star Galaxy Universe
Type IV
1052

1047
Annual Energy Use (J)

Type III
1042

1037
Type II
1032 mit
s i t y Li
ino
Lum
1027
Type I
1022 Earth
1000-year Projections

108 1011 1014 1017 1020 1023 1026

Domain Radius (m)
Figure 1: Log-log plot of annual energy use versus spatial domain radius for a technosphere.
The luminosity limit is defined by the energy use and spatial domain associated with the
extended Kardashev scale [3]. Earth today and projections of Earth’s 1000-year future [4]
all fall below the luminosity limit. Values are provided in Table 1.

4
 lum
use available from stellar luminosity, Emax , and the spatial domain radius,
R:
lum
Emax ≈ R1.5 · 1013.4 J m−1 . (1)
Eq. (1) is not a prescription for how a technosphere must grow but instead is
an upper limit to the energy consumption available within a spatial domain.
Rather than a prediction of future growth, the luminosity limit can be an
instructive tool for thinking about possible future trajectories.
Earth today falls below the luminosity limit because the total energy use
of human civilization is less than the total energy available from sunlight
on Earth’s surface. The spatial domain available to human civilization is
the entire planet, which reflects the ability of human civilization to travel to
and utilize almost any resource on Earth, if desired. This resource domain
could extend in the future, for example if asteroid mining or Mars settlement
become viable. Such possibilities can inform the development of scenarios in
which the technosphere expands beyond Earth.
We consider the set of ten self-consistent projections of Earth’s techno-
sphere 1000 years into the future developed by Haqq-Misra et al. [4], which
are all within the lower left quadrant of Figure 1. These ten scenarios were
developed by conducting a morphological analysis of the poltical, economic,
social, and technological scenario space for Earth’s 1000-year future, per-
forming a cross-consistency analysis to eliminate inconsistent scenarios, and
then applying a novel worldbuilding pipeline to identify the observable prop-
erties of the technosphere. The spatial domain of the scenarios ranges from
Earth alone to the entire solar system, and the range of energy use includes
scenarios both less than and greater than Earth’s energy use today. All ten
scenarios fall below the luminosity limit.
This set of scenarios includes only one that has exceeded the Type I
threshold for annual energy use; however, this case still falls below the lumi-
nosity limit because the domain radius is much larger than the planet. The
domain radius for this and several other cases in the ten scenarios is much
closer to the Type II domain radius, even though energy use is far below
the Type II threshold for annual energy use. In this set of scenarios, annual
energy use increases at a much slower rate with domain radius when com-
pared to the luminosity limit. This suggests that energy use and expansion
may not be as tightly coupled as assumed in the Kardashev [1] scale, and
plausible projections of Earth’s technosphere might be expected to fall below
the luminosity limit rather than to follow it.

5
3. Stellivores: Beyond the Luminosity Limit
The luminosity limit is a hard limit for any technosphere that is based on
harnessing stellar energy within a spatial domain, but such a limit would not
necessarily apply to a technosphere that can extract even larger quantities
of energy from a star. No such cases emerged from the scenario modeling by
Haqq-Misra et al. [4], but the parameter space shown in Figure 1 raises the
possibility of technospheres that might exist above the luminosity limit.
A hypothetical idea by Vidal [5] is that some accreting binary stars might
actually be living or technological systems, with one compact object “feed-
ing” on its companion star to sustain metabolism. Some “stellivores” may
also have higher velocities with trajectories that are directed toward nearby
stars for continued feeding [6]. In such a case a stellivore would be observed
as a binary system not accreting, but ejecting material out of its gravita-
tional well, in order to generate thrust to travel toward another star [7].
Observationally testing the stellivore hypothesis is challenging, but possible,
and Vidal [8] suggested an approach based on characterizing the energy rate
density of such systems. In principle, the idea that some accreting binary
systems could be stellivores is a possibility worth considering as a potential
living or technological system that could exceed the luminosity limit. Ex-
amples of possible stellivore candidates are listed in Table 2 and shown on
Figure 2. These values for accreting white dwarfs, neutron stars, and black
holes are average quantities that were tabulated by Vidal [8].
The spatial domain for stellivores takes two separate cases. First is the
case of “stationary stellivores” that do not necessarily move toward new
feeding targets, in which case the spatial domain is the orbital separation
between the two stars. The stellivore hypothesis proposes three levels of
density: white-dwarf, neutron-star and black hole density. The spatial do-
main for stationary stellivores is chosen as the typical orbital separation for
cataclysmic variables (white dwarfs accreting) and low-mass X-ray binaries
(neutron stars or black holes accreting), which is about 1-2 solar masses or
about 109 meters [9]. Second is the case of “traveling stellivores” that move
over a larger region of space to feed on new stars as needed. In this case,
the spatial domain is a small globular cluster with radius of 10 light years
(∼ 1017 meters), which would provide numerous opportunities for close stellar
encounters.
All stationary stellivores in Figure 2 exceed the luminosity limit. For trav-
eling stellivores in a globular cluster, the accreting neutron star and black

6
 Scenario Annual Energy Use (J) Domain Radius (m) Domain Description
Stationary Stellivores
Accreting White Dwarfs 2.4 × 1038 109 Solar diameter
Accreting Neutron Stars 2.8 × 1040 109 Solar diameter
Accreting Black Holes 2.2 × 1041 109 Solar diameter
Traveling Stellivores
Accreting White Dwarfs 2.4 × 1038 1017 Globular Cluster
Accreting Neutron Stars 2.8 × 1040 1017 Globular Cluster
Accreting Black Holes 2.2 × 1041 1017 Globular Cluster
Star-lifting
Lower Limit for Mass Removal 2.4 × 1031 1011 Earth to Sun
Upper Limit for Mass Removal 2.4 × 1033 1011 Earth to Sun
Planet-like Accretion
Close-orbiting pre-stellivore ∼ 1032 ∼ 109 Radius of Sun

Table 2: Energy use and spatial domain for several stellivore scenarios. Note that Vidal [8]
lists values as energy rate density for an associated stellar mass, which have been converted
here into units of energy per year.

hole cases both exceed the luminosity limit, while the accreting white dwarf
case is very close to the luminosity limit. The idea that a living or technolog-
ical system could extract energy from a star through non-luminous processes,
such as accretion, suggests possibilities for exceeding the luminosity limit.
Other possibilities can be imagined by thinking about the evolutionary
trajectory of stellivores or other technological processes that could operate
above the luminosity limit. A pre-stellivore phase could involve planet-like
accretion that occurs very close to the host star and at a lower energy con-
sumption rate than stationary stellivores, which is above the luminosity limit
with the spatial domain approximately equal to the radius of the Sun. An-
other possibility is the technology of “star-lifting,” which would involve re-
moving sufficient mass to prevent the host star from evolving into a red
giant [10]. Star-lifting would require energy expenditures above the lumi-
nosity limit, if the spatial domain is taken as the Earth-Sun separation.
These hypotheses are worth exploring further in order to make predictions
about civilizations interacting with their star, which might be observable
with present or near-future astronomical facilities.

4. Thermodynamic Efficiency
A major transformation would be required for an Earth-like technosphere
to exceed the luminosity limit, such as in the example of stellivores. We can-
not easily predict the trajectory of such a technosphere, but we can gain

7
 Planet Star Galaxy Universe
Type IV
1052 i t
s s Lim
a
ar M
Sol
1047
Annual Energy Use (J)

Type III
1042

1037 Stationary Stellivores Travelling Stellivores

Star-lifting
Type II
1032
Planet-like accretion? it
i t y Lim
1027 inos
Lum
Type I
1022 Earth
1000-year Projections

108 1011 1014 1017 1020 1023 1026

Domain Radius (m)
Figure 2: Log-log plot of annual energy use versus spatial domain radius for hypothetical
star-eating “stellivores” [5] that extract energy from stars through accretion and could
potentially exceed the luminosity limit. The “star-lifing” cases describe a technosphere
that is prolonging the lifetime of its host star by removing stellar mass [10]. Values are
provided in Table 2. The solar mass limit is defined by the maximum annual energy use
that could be sustained from harvesting mass from a sun-like star.

8
some insight into the necessary transformation by applying a systems ther-
modynamic framework to this problem.
One way of framing the second law of thermodynamics is using the con-
cept of exergy, where exergy refers to the work obtainable within a system.
Using this concept, the second law of thermodynamics states that some “ex-
ergy loss” (i.e., entropy production) will always occur in a system, so that
there can never be perfect efficiency between exergy input to the system (i.e.,
the available useful energy) and exergy output from the system (i.e., prod-
ucts and emissions). The ratio between the exergy outflow and exergy inflow
is the “exergetic efficiency” for a system [see, e.g., 11].
The luminosity limit describes the maximum annual energy use available
from stellar radiation, S, which corresponds to the theoretical maximum ex-
ergy input to a technosphere based on stellar radiation alone. Because of
the second law of thermodynamics, a luminosity-limited technosphere will
never have an exergy output equal to its exergy input, as there will always
be some loss. For a technosphere with annual exergy output X, the exer-
getic efficiency based on stellar luminosity as a source is η = X/S, where
S and X are evaluated at the same domain radius. A luminosity-limited
technosphere is constrained by the limit η < 1. This constraint states that
a luminosity-limited technosphere can never exist exactly at the luminosity
limit but will always experience some inefficiencies that cause exergy loss (or,
equivalently, increase entropy). The theoretical limits of the Type I, II, III,
and IV civilizations on Figures 1 and 2 may therefore be unattainable for
luminosity-limited technospheres.
It is important to note that S and X are annual quantities, tabulated
in units of energy (Joules) per unit time (year) in Tables 1 and 2, which
is an energy flow over time or power. The generation of stellar luminosity
itself is the result of a metabolic-like process: stars release luminous energy
due to the loss of mass that occurs from the fusion of lighter elements into
heavier ones. The total amount of stellar matter lost in this way is very
small compared to the mass of the star, so the luminous energy generated
over the lifetime of a star is only a small fraction of the total energy that
could be available if (hypothetically) the entire stellar mass were converted
into luminous energy. A star like the sun has a typical main sequence lifetime
of about 10 billion years, which will be followed by expansion into a red giant
and eventual evolution into a smaller white dwarf. On this 10 billion year
time scale (and even longer), luminous energy remains an energy flow over
time (power) that is generated by stellar processes ultimately dependent on

9
the mass of the star. The uppermost limit to exergy in a system is therefore
determined by the stellar mass available within the system.
The maximum energy available from stellar mass is E = c2 ΣM , where
ΣM is the total of all mass within the domain radius; this corresponds to
the theoretical maximum exergy input based on stellar mas. This is shown
as the “solar mass limit” on Figure 2, which assumes a total of one solar
mass within the solar system and 1022 solar masses for the galaxy. Two of
the stationary stellivores reside above the solar mass limit (the stationary
accreting neutron star and black hole cases), which indicates that these cases
have an even greater metabolic rate that would require a 100 solar mass or
larger host to sustain its annual energy use. The maximum energy available
from stellar mass is fixed in time (we will limit our consideration to the main
sequence lifetime of a star), so the exergetic efficiency, ϵ, based on stellar
mass as a source is
Xt
ϵ= < 1, (2)
E
where E and X are evaluated at the same domain radius. The annual exergy
output for a particular scenario, X, is an energy flow in time, so Eq. (2) mul-
tiplies X by a specified timescale, t. The values of ϵ relate the exergy outputs
of a civilization over a duration of time to the maximum exergy inputs avail-
able from stellar mass. Stellivores or any other system that harvests energy
from stellar mass may have exergy outputs above the luminosity-based exer-
getic efficiency (i.e., η > 1), but such systems remain bound by mass limits
(i.e., ϵ < 1). If harvesting mass has greater exergetic efficiency than harvest-
ing radiation (i.e., if ϵ > η), then an expanding technosphere may eventually
choose to harness more thermodynamically efficient means of utilizing stellar
resources.
Figure 3 shows the exergetic efficiency based on stellar mass for all sce-
narios considered in Tables 1 and 2, plotted over a 10 billion year timescale.
These calculations follow Eq. (2) and assume that the exergy outputs re-
main constant with time for all scenarios. The luminosity limit is shown as
a shaded blue region comprised of four lines, the top representing Type IV
and the bottom Type I; the increasing trend represents the increased fraction
of luminous exergy outputs over time in proportion to the available exergy
inputs from stellar mass. The trend for Earth is shown in black, falling be-
low the Type I line. All ten of the 1000-year projections (yellow lines) have
a lower exergetic efficiency than present-day Earth, which results from the
larger domain radii of these ten scenarios compared to present-day Earth.

10
 1010
Earth
106 Luminosity Limit
1000-year Projections
102
Solar Mass Limit
Exergetic Efficiency

10 2
10 6
10 10
10 14
10 18
10 22
10 26 0
10 102 104 106 108 1010
Time (yr)
1010
106
100 Solar Mass Limit
102
Solar Mass Limit
Exergetic Efficiency

10 2
10 6
10 10
10 14
Earth
Luminosity Limit
10 18 Stationary Stellivores
Traveling Stellivores
10 22 Planet-like Accretion
Star-lifters
10 26 0
10 102 104 106 108 1010
Time (yr)
Figure 3: Log-log plot of exergetic efficiency over a 10 billion year timescale for the scenar-
ios considered in Table 1 (top panel) and Table 2 (bottom panel). The exergetic efficiency
calculation assumes a solar-mass limit is the maximimum efficiency. Two stationary stel-
livore cases begin above the solar mass limit, and the third exceeds the solar mass limit at
some later time; all of these cases would require a much larger exergy input of 100 solar
masses or greater. 11
The traveling stellivore cases all fall beneath the solar mass limit, with one
case above the luminosity limit. The star-lifter scenario reaches the solar
mass limit by the end of the 10 billion year timescale, which is appropriate
as this is a typical main sequence lifetime for a sun-like star. The stationary
stellivore cases as well as the planet-like accretion cases all eventually exceed
the solar mass limit and even the 100 solar mass limit. These cases would
require much higher host star densities, such as the equivalent of a neutron
star or black hole, in order to sustain the required mass accretion rates.
This framework provides a way to conceptualize different trajectories that
could lead a civilization from conditions like present-day Earth to a stelli-
vore. The possibility space for a civilization’s annual energy use and domain
radius, as shown in Figures 1 and 2, involves a trade-off for a given sce-
nario between exploration by increasing the domain radius and exploitation
by increasing annual energy use [12]. A trajectory of pure exploration would
involve continual expansion of the domain radius with energy use remain-
ing constant, while a trajectory of pure exploitation would involve a fixed
domain radius but a continually increasing energy use. These two extreme
cases are shown as horizontal and vertical vectors in the top panel of Figure
4, along with several other cases. The sub-luminal expansion case involves
both an expansion of domain radius and increase in energy use, but at a rate
that keeps the civilization below the luminosity limit. The late stellivore case
involves a gradually accelerating rate in the increase of energy use, in which
the civilization eventually approaches and may even exceed the luminosity
limit. The early stellivore case features rapid increases in energy use before
the domain radius expands significantly, in which the civilization crosses the
luminosity limit much earlier. The late and early stellivore scenarios suggest
two trajectories between Earth today and a travelling stellivore, only one of
which involves a stationary stellivore as an intermediary state.
The exergetic efficiency of these five trajectories is plotted in the bottom
panel of Figure 4, which assumes that the timescale for the trajectories shown
in the top panel is one million years. The exploration and sub-luminal expan-
sion trajectories both show a continual decrease in exergetic efficiency with
time, as these involve expansion to galactic scales with reduced or no change
to exergetic output. The exploitation trajectory shows a rapid increase in
exergetic efficiency, as domain radius remains constant but exergetic outputs
increase. The late stellivore trajectory shows an initial decrease followed by
an increase in exergetic efficiency and is approaching the luminosity limit.
The early stellivore trajectory crosses the luminosity limit rapidly and ap-

12
 Planet Star Galaxy Universe

1052 it
Lim
ass
o la rM
S Early Stellivore
1047
Annual Energy Use (J)

1042 Exploitation
1037
Late Stellivore
1032
mit
ity Li
1027 inos Sub-luminal Expansion
Lum

1022 Exploration
108 1011 1014 1017 1020 1023 1026
Domain Radius (m)
106
Exploitation
102 Solar Mass Limit
10 2 Early Stellivore
Exergetic Efficiency

10 6
10 10
Late Stellivore
10 14
10 18
10 22 Sub-luminal Expansion
10 26
Exploration
10 30 0
10 102 104 106 108 1010
Time (yr)
Figure 4: Theoretical future trajectories starting from Earth today, shown as annual
energy use versus spatial domain radius (top panel) and exergetic efficiency over a 10
billion year timescale (bottom panel). Colored dots in the top panel correspond to the
scenarios considered in previous figures. The black line and shaded blue region in the
bottom panel correspond to a constant exergy output for Earth and the luminosity limit,
respectively. 13
proaches the solar mass limit before exergetic efficiency begins to reach a
limit or decrease. These five possibilities suggest that different evolutionary
options may be available for civilizations that increase their exergetic effi-
ciency with time compared to those for which exergetic efficiency decreases.

5. Discussion
All of the 1000-year future cases show a lower exergetic efficiency than
Earth today (Fig. 3, top panel) because exergetic efficiency is calculated as a
function of spatial domain, and all these scenarios expand in spatial domain
more rapidly than they increase exergy outputs. This indicates that these
scenarios all emphasize an exploration strategy over exploitation [12], which is
comparable to the exploration and sub-luminal expansion trajectories (Fig.
4). Two of these scenarios (S4 and S7 in Tab. 1) have minimized their
technospheres within Earth’s boundaries, which indicates a focus in these
scenarios on purposefully pursuing other energy use options that are less
energetically efficient; that is, these cases involve an ideological prioritization
to lower exergy outputs over other aspects of their material culture. Other
scenarios involve an expansion of the technosphere across the solar system
(S9 and S10 in Tab. 1), but these still prioritize exploration over exploitation
and show a decrease in exergetic efficiency as the domain radius increases.
The observation that these 1000-year scenarios all show a decrease in
exergetic efficiency may suggest a possible bias underlying these projection
methods or reveal underlying assumptions about expectations for future tra-
jectories. In particular, this is because all ten scenarios feature extant and
non-speciated humans with the Earth’s biosphere still present in some form.
In other words, these scenarios all need to proceed with exploration first
in order to acquire necessary knowledge to ensure that the waste energy
and waste material (and entropy-related degradation in general) is displaced
somewhere else in the domain; otherwise, the civilizations in these scenarios
would overwhelm their spatial domains with waste or similarly deplete or
degrade them. It is possible that following an exploitation trajectory imme-
diately might destroy the host planet’s biosphere, so such a trajectory might
be viable only after establishing an uninhabited industrial “service world”
[13]—for example on Mercury, in the case of the Earth’s biosphere, so as
not to risk runaway climate change or any other form of degradation that
could lead to a biospheric collapse. This problem is related to unavoidable

14
degradation due to absolute entropy increases in smaller domains (including
on Earth today), which likewise may be harmful.
The early stellivore trajectory (Fig. 4) involves a transition from Earth
today through a planetary accretion phase, followed by a stationary stellivore,
and then a traveling stellivore. We can imagine a generalized narrative that
might explain such a trajectory. In phase 1, the exponential energy demands
of computing leads to a gradual heating up of Earth, eventually transforming
it into a fusion powered (post)planet. This is a temporary solution, as the
mass of a planet is tiny in comparison to a star, so even using almost all
the atoms of Earth as fuel for fusion reactors would eventually run out much
more quickly than tapping directly into the host star. Phase 2 would reach
the point of planetary accretion or star-lifting, as these strategies support
each other: a civilization could extract mass from the host star, thereby
prolonging its lifetime, and then use this mass as a new kind of energy source
beyond stellar radiation (i.e., beyond the luminosity limit). At phase 3, the
civilization begins the transition to a traveling stellivore, as the host star will
eventually run out and so another nearby star will be needed. This leads to
phase 4, in which the civilization begins accreting another star, and the cycle
repeats, with the civilization even pursuing “hardware density transitions”
to expand beyond main sequence-type stars to instead sustain metabolism
on denser white dwarf, neutron star, or black hole hosts. The extent to
which a civilization could persist as a stellivore, particularly one that feeds
on large amounts of mass, depends on the availability of sufficient sources of
accretable mass to sustain the necessary metabolic rate. With Earth today
as a starting point, this early stellivore trajectory would require some form
of complete unification and centralization in order to sacrifice the biosphere
for the sake of a merged-transhumanity, with a singular vast exergy flow
that sustains the entire civilization. By way of metaphor, the early stellivore
trajectory is akin to a “hunter-gatherer” approach toward resource collection;
alternatively, this is analogous to the ecology of predation, namely operating
like true predators or parasitoids depending on how fast they kill their host
star.
The late stellivore trajectory (Fig. 4) does not include a planetary accre-
tion or a stationary stellivore phase, but instead this trajectory remains below
the luminosity limit until the civilization’s domain radius has reached galactic
scales. The exergetic efficiency for the late stellivore trajectory begins with
a decrease, comparable with the sub-luminal expansion trajectory, but then
begins to increase and approach the luminosoty limit. This suggests an alter-

15
native path toward becoming a travelling stellivore. One possibility could be
that some of the 1000-year scenarios could be consistent with the early stages
of a civilization that later becomes a traveling stellivore. Another possibility
is that none of these 1000-year scenarios have any logical continuity with a
future as a traveling stellivore, but that some other late stellivore trajectory
might exist that begins with Earth today. Still another possibility is that no
viable late stellivore trajectory exists that includes Earth as a starting point.
By way of metaphor, the late stellivore trajectory can be compared to the
approach of a “farmer” toward resource acquisition, or from an ecology of
predation viewpoint, a grazer or a parasite which does not completely kill
its prey (in this case, the host star). The late stellivore trajectory may not
necessarily require centralization and unification and instead might involve
multiple instances or “back-up nodes” of a technosphere that is farming ex-
ergy across its domain. Such efforts would attempt to ensure that “recovery
copies” of the technosphere exist in multiple places (i.e., different host stars)
before attempting to cross the luminosity limit at any one of these stars to
create a traveling stellivore, while the distributed civilization can continue
to utilize the existing infrastructure around other stars that is already in
place. If the domain of a single successfully created traveling stellivore is the
entire galaxy, thanks to its precursor civilization, then even when it crosses
the luminosity limit of a single host star, its overall place in Figure 4 (top
panel) would be below the luminosity limit point for the whole galaxy—at
least until it eventually takes advantage of the infrastructure its civilization
has distributed around other stars and starts feeding on them.

6. Conclusion
The exploitation-exploration framework can provide useful insights into
thinking about the possible trajectories of Earth’s future and advanced tech-
nospheres, even very long-lived technospheres that are vastly different than
our own. The 1000-year future scenarios illustrate cases closer to pure explo-
ration (including sub-luminal expansion) that are consistent with plausible
projections of Earth’s technosphere today, given that humans have not speci-
ated in any sense. By contrast, both the early and late stellivore trajectories
represent versions of the multi-phased optimal solution to the exploitation-
exploration trade-off, depending on the life-stage of a technosphere and its
environmental conditions. The framework developed by Berger-Tal et al. [12]
(see, e.g., their figures 1 and 2) suggests four phases of establishment, ac-

16
cumulation, maintenance, and exploitation with varying durations; however,
not all four of these phases are required, and some particular versions of the
optimal solution could even involve only the two phases of establishment and
exploitation for subjects with very short life-spans. It is worth considering
the extent to which the trajectories of advanced or exotic technospheres in
particular environments might follow different versions of such a multi-phased
solution that optimizes their exploration-exploitation strategies in ways that
might be inaccessible to Earth’s technosphere today.
Ultimately, any hypothetical exploration of civilizational trajectories, no
matter how exotic or mundane, should lead to observational predictions if
they are to be useful in aiding the search for technosignatures. One impor-
tant lesson from this study is to recognize that luminosity and mass both
serve as physical limits to growth, and a civilization’s strategy in maximiz-
ing its expansion in spatial scale versus its increases in energy use can lead
to different observable consequences. The 1000-year scenarios that fall close
along the pure exploration trajectory may prove to be among the most diffi-
cult to detect—or, at least, discovering such a civilization may require more
serendipitous conditions when compared to others that expand in both spa-
tial scale and energy use. But the trajectories that are closer toward an
optimal balance between exploration and exploitation also raise challenges
in observation, given that such technospheres may be vastly different than
that on Earth, and so recognizing their associated technosignatures may not
be obvious. For stellivores, possible approaches toward testing this hypoth-
esis include examining candidate traveling stellivore for possible signatures
of a stellar engine [7], testing the goal-directedness of candidate traveling
stellivore [6], testing whether accretion in candidate stationary stellivores
is controlled, and applying biological metabolic scaling laws to candidate
stellivores [e.g., 8]. Such efforts would be an important first step toward un-
derstanding the extent to which known stellar systems could reveal unusual
properties that are closer to living systems than they first appear.
In summary, this paper calls for more broad and imaginative thinking
about the possible trajectories that could lead to a remotely detectable tech-
nosphere. Previous attempts at searching for characteristic signs of Type
I, II, III, and IV civilizations may inadvertently presume that an advanced
civilization will develop along the luminosity limit as a trajectory, but achiev-
ing the luminosity limit itself remains out of bounds for a luminosity-limited
technosphere due to thermodynamic constraints. Likewise, any civilization
that can exist at or beyond the luminosity limit will necessarily have dis-

17
covered a way to derive greater exergetic efficiency from its environment,
such as by harvesting stellar mass directly instead of relying only on stellar
luminosity.

Acknowledgments
This article was inspired by the session on “Advancing the Search for
Technosignatures” and other conversations at the 2024 Astrobiology Science
Conference in Providence, Rhode Island. J.H.M. gratefully acknowledges
support from the NASA Exobiology program under grant 80NSSC22K1009.
Any opinions, findings, and conclusions or recommendations expressed in
this material are those of the authors and do not necessarily reflect the views
of their employers or NASA.

References
[1] N. S. Kardashev, Transmission of information by extraterrestrial civi-
lizations, Soviet Astronomy 8 (1964) 217.

[2] M. Cirkovic, Kardashev’s classification at 50+: A fine vehicle with room

for improvement, Serbian Astronomical Journal (2015) 1–16. doi:10.
2298/SAJ1591001C.

[3] R. H. Gray, The extended Kardashev scale, The Astronomical Journal

159 (2020) 228. doi:10.3847/1538-3881/ab792b.

[4] J. D. Haqq-Misra, G. Profitiliotis, R. Kopparapu, Projecting earth’s

technosphere. I. Scenario modeling, worldbuilding, and overview of re-
motely detectable technosignatures., Technological Forecasting and So-
cial Change (in review). doi:10.48550/arXiv.2409.00067.

[5] C. Vidal, Stellivore extraterrestrials? Binary stars as living systems,

Acta Astronautica 128 (2016) 251–256. doi:10.1016/j.actaastro.
2016.06.038.

[6] C. Vidal, A. Nilipour, J. Hoang, O. A. Johnson, Are spider pulsars goal-

directed? Tests of the stellivore hypothesis with Gaia DR3, in: Penn
State SETI Symposium 2023., 2023.

18
 [7] C. Vidal, The spider stellar engine: A fully steerable extraterrestrial
design?, Journal of the British Interplanetary Society 77 (2024). doi:10.
59332/jbis-077-04-0119.

[8] C. Vidal, Energy rate density as a technosignature: The case for stel-
livores, in: Life in the Universe: Big History, SETI and the Future
of Humankind. IBHA & INAF-IASF MI Symposium, 2019, pp. 15–16.
doi:10.5281/zenodo.14007941.

[9] R. W. Hilditch, An Introduction to Close Binary Stars, Cambridge Uni-

versity Press, 2001.

[10] M. T. Scoggins, D. Kipping, Lazarus stars: Numerical investigations of

stellar evolution with star-lifting as a life extension strategy, Monthly
Notices of the Royal Astronomical Society 523 (2023) 3251–3257. doi:10.
1093/mnras/stad1617.

[11] S. Huysman, T. Schaubroeck, J. Dewulf, Exergy and cumulative

exergy use analysis, Sustainability Assessment of Renewables-Based
Products: Methods and Case Studies (2015) 141–154. doi:10.1002/
9781118933916.ch10.

[12] O. Berger-Tal, J. Nathan, E. Meron, D. Saltz, The exploration-

exploitation dilemma: A multidisciplinary framework, PloS one 9 (2014)
e95693. doi:10.1371/journal.pone.0095693.

[13] J. T. Wright, J. Haqq-Misra, A. Frank, R. Kopparapu, M. Lingam,

S. Z. Sheikh, The case for technosignatures: Why they may be abun-
dant, long-lived, highly detectable, and unambiguous, The Astrophysi-
cal Journal Letters 927 (2022) L30. doi:10.3847/2041-8213/ac5824.

19