Five Things to Look For at AWS re:Invent 2024
The turkey has been eaten and the discount TVs have been purchased, which signals that it’s time to partake of the next time-honored tradition in the holiday schedule: Attending AWS re:Invent in Las Vegas.
In 2023, Approximately 65,000 people attended re:Invent, making it one of the largest technology conferences on the schedule. The conference this year is expected to attract a similar number of people for about 2,500 sessions spread across a dozen or so casinos along the Strip.
AWS’s annual shindig is not only a great place to learn about how to maximize the public cloud service, but also to learn about vendor solutions. About 2,000 companies exhibited in the expo last year, including hundreds of vendors hawking various big data management, advanced analytics, and AI offerings.
As re:Invent 2024 gets going, here are five things to keep an eye out for.
1. New AWS CEO Matt Garman
This will be the first re:Invent as CEO for Matt Garman, who was hired to replace Adam Selipsky in June. But it’s not the first re:Invent for Garman, who has worked for AWS since 2006. As an AWS insider and an AWS lifer, Garman is being counted upon to deliver a new phase of growth for AWS in ways that Selipsky did not.
As is custom, Garman will deliver the first executive keynote, scheduled for Tuesday morning at 8 a.m. PT; you can watch livestreams of all of the keynotes here. This will be the first time that many AWS customers and partners get to hear from Garman himself, and so it will be the first impression for many.
What sort of style does Garman have? Before he left for the top job in 2021, Andy Jassy was famous for delivering long, fun keynote addresses. In his short three-year tenure as AWS CEO, Adam Selipsky, who was hired from Salesforce and Tableau (although he worked at AWS before), delivered keynotes that were more or less forgettable.
Garman, who Jassy reportedly considered for the top job before picking the outsider Selipsky to essentially hold down the fort during Covid, has been described as a “wartime CEO” for AWS. It will be interesting to see how this plays in front of the troops in Vegas.
2. Low-Level Services or High-Level Apps?
AWS is famously tight-lipped about products under development, so there’s no telling what the tri of top AWS executives–Garman on Tuesday, VP of AI Swami Sivasubramanian on Wednesday, and CTO Werner Vogels on Thursday–have up their collective sleeves. Undoubtedly, there will be surprises.
One important question is whether AWS’s strategy will shift under Garman. The company currently offers about 200 low-level services, according to this AWS white paper. Will Garman continue that strategy, or perhaps his ascension signals a sea-change at AWS and a move toward delivering more full-fledged applications?
AWS does deliver some applications, of course, but it has appeared been reluctant to roll out more SaaS applications that would compete with its partners. However, revenue growth has been a concern over the past three years under Selpisky, so perhaps there will be some changes (or perhaps not).
3. Show Progress in GenAI
Ever since OpenAI dropped ChatGPT on the world smack dab in the middle of re:Invent 2022, the momentum behind generative AI has been growing. AWS has responded to this market shift in several ways, including by ramping up availability of Nvidia GPU instances as well as its own Inferentia and Trainium chips; launching Amazon Q, its GenAI-powered AI assistant; and delivering LLM checkpointing for PyTorch, among many other moves.
On the vector database front, AWS has added vector capabilities to many of its databases, enabling them to play roles in retrieval-augmented generation (RAG) GenAI workflows. It has also combined vector search and graph capabilities. What AWS lacks is its own native vector database, which we pointed out last year. AWS so far has stuck with hybrid vector capabilities in existing databases, but will that cut the mustard as GenAI workloads scale up? Only time will tell.
More broadly, AWS needs to show that it’s moving quickly and decisively in GenAI as it fights a perception that it trails its two public cloud competitors, Microsoft Azure and Google Cloud, in terms of GenAI capabilities. The company has invested millions in an ambitious plan to train its own large language model (LLM), called Olympus, which will reportedly go beyond its current Titan line of LLMs, but so far, Olympus hasn’t been launched. The departure of Matt Wood, the former VP of AI, to PwC last month also doesn’t help.
4. Demonstrate Leadership in Agentic AI
The AI space has moved quickly from the chatbot and copilot phase directly into semi- or fully-autonomous agents, or agentic AI. This represents an opportunity for AWS to show market leadership.
AWS offers some semi-autonomous agents via Amazon Q, which it unveiled last year. In May, the company bolstered Q with new Developer Agents, which can essentially function as junior-level programmers or business analysts. The company has been vocal about the potential for Q to automate rote tasks, such as writing boilerplate Java code, and now it will need to take it to the next level.
As competitors Salesforce, Microsoft, and Anthropic make inroads into agentic AI, AWS will be under the gun to keep in front of this next wave of AI. With the world’s largest cloud platform and some of the best AI developers in the world, the potential for great things is there, but it will be up to AWS leadership to make it happen.
5. Impact of Return-to-Office
There will be one big wild card looming like the Sphere in Vegas? How will Andy Jassy’s company-wide return-to-office requirement impact morale and productivity at the largest cloud computing company?
Jassy, who in 2021 went from AWS CEO to succeeding Jeff Bezos in the top job at the online bookseller, issued his edict in September that all Amazon employees return to the office five days a week starting in 2025.
That requirement has not been sitting well with some of Amazon’s 1.5 million employees. Fortune reported that Amazon employees are “rage applying” for other jobs, and speculated that Jassy’s requirement could have the same impact as layoffs. AWS has let 27,000 employees go over the past three years under two rounds of layoffs under Selipsky, and this could be in effect the third round.
Related Items:
The Semi-Autonomous Agents of Amazon Q
10 Things You Need to Know About Matt Garman, the Incoming CEO of AWS