Amazon v. Brian Hall - Declaration of Matt Wood, AWS VP of AI
Amazon v. Brian Hall - Declaration of Matt Wood, AWS VP of AI
Amazon v. Brian Hall - Declaration of Matt Wood, AWS VP of AI
16 1. Identity of Declarant. I have worked for Amazon Web Services, Inc. (“AWS”)
17 for over 10 years. My current title is Vice President of Artificial Intelligence, a position I have
18 held since September 2019. Before I became AWS’s Vice President of Artificial Intelligence, I
19 was its General Manager of Deep Learning and Artificial Intelligence. I held that position from
20 January 2017 until September 2019. Before that, I was AWS’s General Manager of Product
21 Strategy, which position I held from October 2012 until January 2017, where I acted as a script
24 issues that give me familiarity with and access to Amazon business records and emails. I am
25 very familiar with Brian Hall’s work for Amazon from when he joined Amazon in June 2018
26 until his resignation in February 2020. We worked together on numerous projects and
27
2 indicated, all facts set forth in this declaration are based on my personal knowledge, which is
3 based on my experience and my review of relevant documents. The facts and documents set
4 forth below are maintained in the ordinary course of business at or near the time of the act,
5 condition, or event. If called upon to testify regarding such matters, I could competently do so.
6 BACKGROUND
8 Machine Learning (“ML”) efforts, described below. I partner closely with all the AWS
9 stakeholder teams for ML cloud, such as finance, HR, partner, sales, business development, and
12 computer to improve its ability to perform tasks and to mimic the ways by which humans learn,
18 technologies into their own businesses for use on their own data sets. Given the breadth of its
20 business and revenue opportunity for Amazon. ML has emerged as a beachhead as to how
21 Amazon customers choose their cloud provider (i.e., AWS, Google Cloud, or Microsoft).
22 6. AWS ML’s Competition with Google ML. AWS and Google provide ML cloud
23 services that compete head-to-head in a market that is growing by billions of dollars per
24 year. For example, like AWS, Google provides a host of pre-programmed cloud ML
25 applications, including those related to computer vision and speech recognition, in direct
26 competition with AWS. AWS competes with Google for ML thought leadership and mind
27
2 Studies show that customers often choose their cloud provider primarily based on that
3 provider’s ML capabilities. Mr. Hall was a crucial part of AWS’s discussions regarding what
4 and where to push AWS’s capabilities forward, why and for how long to defer where AWS had
5 opportunities for improvement, and how to develop AWS’s thought leadership in ML,
7 7. Declarant’s Relationship with Mr. Hall. In June 2018, Mr. Hall became
8 AWS’s Vice President of Product Marketing. From June 2018 to February 2020, I worked
9 continuously and closely with Mr. Hall. Among other things, I worked closely with Mr. Hall in
11 8. As with all AWS products, Mr. Hall was responsible for strategically developing
12 the “story” around AWS’s ML products prior to their launch. Mr. Hall and I worked together
13 on marketing strategies for numerous AWS ML products that AWS has yet to launch and
14 deciding where marketing efforts would be most effective in driving AWS’s message and the
15 product’s adoption. By virtue of Mr. Hall’s involvement in creating related marketing plans,
16 Mr. Hall possesses significant amounts of highly confidential non-pubic information regarding
17 these yet-to-be launched AWS ML products. This includes technical specifications, intended
18 customers, messaging strategies, pricing strategies, and anticipated launch dates; the amount of
19 marketing effort AWS anticipates is necessary; and which customers have requested such
20 products.
22 9. Mr. Hall also led development of the overall product marketing strategy for
23 AWS ML, including but not limited to his role in: (a) analyzing AWS’s operational planning
24 documentation (“OP1”), which sets out the strategy for AWS’s ML business through the end of
25 2020; (b) participating in weekly business reviews of key metrics and status reports on AWS’s
26 overall ML strategy; and (c) analyzing monthly update reports regarding the status and
27
2 reports to AWS’s CEO, Andy Jassy. Examples of business reviews in which Mr. Hall
3 participated are attached as exhibits. Exhibit A is a true and correct copy of an AI Weekly
4 Business Review (“WBR”) Summary; Exhibit B are true and correct copies of AWS’s
5 December 2019 and February 2020 ML Marketing Monthly Business Review (“MBR”)
6 Summaries.
8 is reflected in Exhibit C, which is a true and correct copy of a February 5, 2020 email that both
9 Mr. Hall and I received regarding a high-level ML strategy meeting to discuss AWS’s overall
10 business plan for AWS ML in 2020. Exhibit D is a true and correct copy of the highly
11 confidential 2020 AWS Machine Learning Category Market Plan attached to that email. This
12 document contains extremely sensitive AWS information, including AWS’s customer analysis,
13 AWS’s strategic business goals for 2020, specific AWS initiatives to be launched in 2020, and
16 11. Mr. Hall’s Role in Competitive Analysis of Google. A large part of Mr. Hall’s
17 role with AWS ML was product positioning—i.e. ensuring AWS’s ML products were
19
20
21
22
23
24
25
26
27
2 in pushing AWS forward as an ML innovator. Microsoft and Google both copied this effort –
3 the day after AWS’s announcement of DeepLens, Google even announced that it would soon
4 have a camera available too, but the damage was done and AWS was at the forefront. With
10
11
12
13 .
14 13. Mr. Hall’s team constantly worked on pushing my team to innovate beyond
15 what was available commercially and made it a priority to stay ahead of Google in marketing
16 efforts. During the course of Mr. Hall’s work on marketing plans for these products, he learned
17 large amounts of highly confidential and proprietary information regarding the products and
18 associated marketing strategies. I describe three examples of these efforts, set to launch this
19 year and into 2021, below. These are just are three examples out of more than
27
7 AWS’s annual conference for all things cloud computing, Mr. Jassy devoted a half hour to
8 SageMaker.
9 16. Because of its success, many companies, including Google, attempt to imitate
10 SageMaker and to stay ahead, AWS must have a very clear vision of how to continue to
13 AWS’s roadmap for SageMaker. Attached as Exhibit F is a true and correct copy of an
14 October 28, 2019 email that both Mr. Hall and I received regarding a meeting to discuss
15 SageMaker attaching two highly confidential documents which provide a comprehensive list of
16
18 its messaging regarding competing products to these AWS customers in order to undercut
20 17. Most recently, one of the “stories” that Mr. Hall helped AWS develop was
21 regarding how
22
23
24
25
26
27
2 below.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
10
11
12
13
14
15
16
17
MR. HALL’S EMPLOYMENT WITH GOOGLE
18 WILL ENABLE GOOGLE TO COMPETE UNFAIRLY WITH AMAZON
19 21. Mr. Hall’s Proposed Temporary Position Working for Google Cloud Next. I
20 have reviewed Google’s proposal that Mr. Hall work on Google Cloud Next. Google Cloud
21 Next is Google’s premiere cloud product marketing event of the year and where Google makes
22 key announcements regarding its capabilities, positioning (such as what releases it has in beta
23 or pre-release and what is generally available), and targeting. My team watches Google Cloud
24 Next incredibly carefully as this conference has the potential to upstage our anticipated
25 launches and interrupt our positive press for the next year. I am especially sensitive to the
26 impact of Google Cloud Next given my previous work in product strategy and speech writing
27
2 concrete impact on the business due to a shift in perception and press coverage. If Mr. Hall
7 .
8 22. Further, there is no way to prevent Mr. Hall from doing so. Even the most
9 conscientious of us must draw on the full extent of our knowledge and experiences in drafting,
10 revising, and editing speeches that will drive the vision of a cloud computing company. This is
11 equally true with Mr. Hall providing play-by-plays of what Google announces. Mr. Hall would
12 essentially be crafting sound bites for the press based on his judgment and determination of
13 what is important. For example, Mr. Hall knows the weaknesses and strengths of AWS’s yet to
14 be launched ML products. He could influence thought leadership (such as major customers and
15 analysts) into believing areas where AWS is anticipating launching a key product are not as
16 important as the areas where Google has an offering, knowing areas where AWS is not
17 planning a competing product launch. He has the benefit of AWS’s confidential roadmap,
18 strategy, narratives, and launch information. Mr. Hall’s editorial decisions cannot be distanced
19 from the two years Mr. Hall spent determining this strategy on behalf of AWS nor the many
20 months he specifically spent working on re:Invent, AWS’s own version of Cloud Next.
22 reviewed Google’s letter describing Mr. Hall’s proposed role at Google as VP of Product
23 Marketing for Cloud. This role does not seem to be distinguished any meaningful ways from
24 Mr. Hall’s work at Amazon. Google proposes that Mr. Hall have immediate responsibility for
25 the narratives and marketing strategy for Google’s brand and products, working with analysts,
26 and creating public facing materials. This position competes directly with the position Mr. Hall
27
2 knowledge of AWS’s plans to advance its thought leadership and mind share as well as his
3 knowledge of AWS’s years of development toward ML product suite “stories,” and his
4 detailed, comprehensive, and highly confidential knowledge regarding over a dozen anticipated
5 launches of AWS products over the next two years would provide Mr. Hall a roadmap for how
7 Google could also use the information in Mr. Hall’s possession to spur Google’s development
8 of competing Google products. Similarly, Google could use this information to launch
9 competitive marketing messaging before AWS actually launches its cloud products according
10 to the cloud roadmap and marketing plans Mr. Hall developed for AWS before leaving to work
11 for Google. Again, the competition between Amazon and Google is fierce, and even the
12 slightest inference (or enhancement or deviation in a pre-drafted narrative) from a former AWS
13 Vice President who was just recently deeply immersed in AWS’s roadmap would undoubtedly
14 (and unfairly) affect a customer’s decision to migrate workloads to Google over AWS. Hall, as
15 Vice President of Product Marketing for Google, will necessarily make this his primary
16 objective.
17
18 I declare under penalty of perjury of the laws of the state of Washington that the
20
EXECUTED at ____Seattle__________ this ___8th__ day of June, 2020
21
22
23
24
25
26
27