There has been explosive growth in modern data stack in the last 2 years. At Gemba Capital, we have mapped out the data stack landscape – new categories, tools, trends, and opportunities. 👇🧵

Looking ahead, AI holds immense potential and its right time to build. Please share your thoughts and comments. If you’re a founder building in GenAI, pls do reach out to me - kamini@gembacapital.in

Meta’s earnings call provided tangible evidence that AI investments deliver measurable value, even at such large scale (4Bn MAUs) Blog👉

Big Tech firms have leveraged AI to enhance core products. Referencing to my last blog- Origins of GenAI which highlighted how GenAI took birth inside labs of Alphabet and Meta and how it powers personalised recommendations in Insta and Youtube.

AI is beginning to drive earnings growth, as seen in strong performance in Google and Microsoft’s cloud computing and improved advertising margins at Meta.

What’s driving AI spend? Is this all madness and hype? OR Can meaningful ROI be generated over long term? Proof is in the pudding !!

This is an arms race, with each company striving to control the entire GenAI stack—from chip design and GPU manufacturing to model development and the application layer—rather than just own a piece of supply chain.

📌 Where AI spend is going? As AI gains mainstream adoption, Big Tech firms are investing heavily not just in model development, cutting $Mn dollar paychecks for AI researchers but focusing on building physical infrastructure for AI compute - data centers, GPUs,silicon chips

Big Tech firms hold substantial cash reserves. Their free cash flow is declining—Alphabet’s is down to $5.3 Bn and Meta’s to $4.2 Bn in Q2’25 primarily due to aggressive AI investments. Whether these massive spends will yield tangible returns remains an open question

Since AI gained mainstream focus in 2022, capex spending has nearly tripled for these firms. About 60% of each firm’s capital investments are now directed toward AI infrastructure, a significant shift from previous years when emphasis was on mobile services and software R&D.

📌 BigTech Capex spend surges - ‘When you talk about AI, you talk in Billions and not Millions’ ▶️ Big Tech firms are collectively expected to allocate approx $350 Bn in capex in 2025, increasing to around $400 Bn in 2026. This suggests that each co capex budget exceeds $100 Bn

I have been tracking capex by Big Tech firms, as it serves as a proxy for AI investment. This approach enables me to build out thesis on emerging AI trends, the evolution of core AI infra, and how this will power new business models, applications in GenAI era 💡

📣 Last week, BigTech firms (Alphabet, Microsoft, Meta, and Amazon) reported robust Q2 earnings, exceeding analyst expectations with double-digit growth in revenue and earnings. Their strong performance boosted market cap, with Microsoft surpassing a $4 Tn and Meta ~ $2 Tn 👏

AI Infra Arms Race: Big Tech’s Billion-Dollar Bet on the Future #bigtech #earnings #meta #google #openai #genai #venturecapital #entrepreneruship #buildinginai @GembaCapital @Adithpodhar

#twitteracqusition, #founderrights #entrepreneurship #elonmusk #socialnetwork #decentralisation

The book traces lofty goals, unrealistic targets and the pivotal role of founder rights, set against constant comparisons with Meta and Google. Ultimately, both Musk and Dorsey realising that power and control over global speech is much greater burden.

Overall I enjoyed reading the book. "Battle for the Bird" offers a gripping insider look at Twitter’s turbulent transition to ‘X’ after Elon Musk’s 2022 takeover, including his spontaneous and haphazard decision was to buy twitter. Eventually he overpaid for the deal

While Twitter didn’t take this route, Jack Dorsey later launched Bluesky - open, decentralized social network. Will we see a future where social networks offer real choice, control over algorithms, & interoperability, letting us join communities that truly reflect who we are?

Content moderation, speech, and control wouldn’t be dictated by a handful of execs, but shaped by the communities themselves. The idea of a decentralized protocol powering multiple social networks one that puts users in control is compelling.

If social networks operated as open protocols, each company could use the underlying tech to purpose build networks with their own unique designs, policies, and algorithms. Networks with extremist ideologies can exist as well.

Imagine if Twitter(X) and other social networks had been built as an open internet protocols, like email/ HTTP. With email, every provider (Yahoo, Google, etc.) offers its own service, but the underlying protocol allows interoperability. You can send an email from Gmail to Yahoo.