Good morning. I’ve got one more week in chilly Jersey weather until HumanX, a big old AI conference out in Las Vegas.

I’ll be there, attending (and even moderating a few panels). Will be diving deep into AI integrations in healthcare and the enterprise. Should be very interesting.

If anyone plans to be there, shoot me a note!

— Ian Krietzberg, Editor-in-Chief, The Deep View

In today’s newsletter:

• 🌎 AI for Good: Saving the whale sharks

• 💰 As AI Trade shakes, TSMC gears up for $100 billion US investment 

• 📱 Apple is slow-playing a generative AI-enhanced Siri

AI for Good: Saving the whale sharks

Whale sharks, those massive — though harmless — bus-sized sharks, are, like many marine species, endangered. Though exact population numbers for whale sharks aren’t known, the International Union for Conservation of Nature, due to the impact of ship strikes, pollution and fishing, has declared the species endangered. 

And, like many conservation efforts, the conservation group Shark Team One is working on a project that leverages machine learning technology and photography to better protect whale sharks. 

• The project aims to automatically classify images of injured whale sharks using machine learning to create a vast, detailed catalog of ship strikes and injured sharks. 

• Broadly, the idea is to provide information to scientists, conservationists and policymakers to better inform policies and approaches that can actually protect the sharks. 

Why it matters: More specifically, the project aims to construct a “wildlife corridor” for the whale sharks using the data parsed and processed by the team’s machine learning tech, data that can be provided to ship captains to ensure that vessels avoid areas where they might hit a shark.

Get Your AI Voice Agent Up and Running in Minutes with Vapi

Voice AI is at a tipping point—and Vapi is leading the revolution. Whether you need a simple inbound voice agent or a complex, fully interactive AI-driven solution, Vapi is the go-to platform. Use any LLM and voice model, connect your tools, write your prompt—and give it a call!

Join Over 100,000 Developers—from disruptive startups to Fortune 500 companies—who trust Vapi as the most configurable API for building, testing, and deploying voice agents at scale.

Why Vapi?

• Flexible integration with your own data, APIs, and models

• Human-like agents for natural conversations

• Scale to millions of calls at <500ms latency

• Enterprise-grade security with LLM guardrails with HIPAA/SOC-2

Ready to try it?
Call 1‑844‑439‑8274 or click here to talk to Vapi.

As AI Trade shakes, TSMC gears up for $100 billion US investment

The news: Taiwan Semiconductor Manufacturing Co., or TSMC, the manufacturer that powers Nvidia’s GPU supply chain, on Monday committed to investing an additional $100 billion into its U.S.-based chip work. 

The details: Announced by President Donald Trump — and somewhat shortly following Trump’s announcement of Project Stargate, a $500 billion AI data center project spearheaded by Softbank and OpenAI — the investment will include the construction of three new fabrication centers, two advanced packaging facilities and a major research and development facility. It brings TSMC’s total U.S. investment up to $165 billion. 

• TSMC said that it expects the move to bring tens of thousands of jobs to the local economy, including those in construction and chip manufacturing. The facilities, like TSMC’s previous U.S.-based facilities, will be built in Arizona. 

• TSMC has received as much as $6.6 billion in grants from the Biden-era Chips Act, a piece of legislation designed to bring semiconductors back to U.S. soil. 

The landscape: The news coincides with Trump’s final affirmation that 25% tariffs will be instated against Canada and Mexico starting today, a confirmation that, though expected, pulled the markets down. 

The Nasdaq slid more than 2.6%, the S&P 500 slid more than 1.5% and Nvidia slipped a full 8% on Monday. Nvidia is now down 15% for the year, and is around 25% off its 52-week high of $153, which was achieved in early January. TSMC, meanwhile, retreated more than 4%. 

“Whether the stock market can survive this change remains to be seen,” Chris Rupkey, chief economist at FwdBonds, wrote. “One way or another, tariffs will be a shock for the economy.”

This tough start to March trading — despite TSMC’s announcement — follows a red February in which early signs began to appear that the AI trade, prodded along by macroeconomic and geopolitical factors, is beginning to unwind. 

• AI funding: Anthropic, the AI startup behind Claude, closed a $3.5 billion funding round at a $61.5 billion post-money valuation. So, closer to OpenAI, but still not quite there — OpenAI is currently in talks to raise at a $260 billion valuation.

• AI phone: T-Mobile’s parent company launched a new “AI phone” in partnership with AI Search startup Perplexity on Monday. A number of GenAI apps will be baked into the phone, with Perplexity’s GenAI assistant acting as a Siri-like assistant. The phone will be released sometime this year.

• YouTube tries again to compete with Amazon to become all-video hub (The Information).

• UK, France scramble to draft peace plan for Ukraine as US support falters (Semafor).

• China wants tech companies to monetize data, but few are buying in (Rest of World).

• Amazon’s delivery drones are grounded. The birds and dogs of this Texas town are grateful (Wired).

• Nvidia’s unofficial exports to China face scrutiny after arrest of silicon smugglers in Singapore (CNBC).

Apple is slow-playing a generative AI-enhanced Siri

Apple, who’s been pretty hush-hush on generative AI since the Race began, is reportedly locked up in further delays to upgrade Siri with a dose of GenAI enhancements. 

Bloomberg’s Mark Gurman reported over the weekend that people within Apple’s AI division believe that a “true modernized, conversational version of Siri won’t reach consumers until IOS 20 at best in 2027.” 

• Gurman said that Apple does plan to ship a Siri update later this year, likely at WWDC in May, one that will finally combine all those Apple Intelligence features we’ve been hearing about into a better version of Siri. 

• But he affirmed that a Siri that would be on par with the latest chatbots isn’t coming anytime soon. This sudden Siri lag comes after the IOS assistant popularized the very idea of chatbots more than a decade ago.

“People involved in Apple’s AI work say its foundational and large language models — the basis for its homegrown AI features — are reaching their limits,” Gurman wrote. “There also have been problems with rivals poaching talent … getting its hands on enough chips has been another challenge.” 

Gurman’s report shortly follows Amazon’s announcement of Alexa Plus, an LLM-enhanced upgrade of the system that pointedly offers custom personalization; the idea is that Alexa will soon be easier to talk to and far more functional, acting as the thread that connects digital ecosystems, including personal data, accounts and smart home tech, etc. 

“We continue to view this first launch of Apple Intelligence as just the beginning of a broader AI strategy for Apple as we estimate roughly 25% of the world's population will eventually access AI through an Apple device over the next few years,” Wedbush’s Dan Ives wrote in a Monday morning note.

“Of course, Apple is late to AI but they are essentially a toll collector on its unmatched global ecosystem as any AI app to consumers will ultimately go through Apple as we saw firsthand with DeepSeek last month. This continues to be the major piece of the Apple AI strategy that investors are missing … it’s less about the killer app/LLM from Apple itself and more around being a foundation for consumer AI agnostic of the LLM.”

Shares of Apple fell 1.5% Monday.

With a market valuation of $3.5 trillion, Apple is the largest publicly traded company in the world by a large margin; Microsoft, with a valuation of $2.9 trillion is in the number two spot, about even with Nvidia.

Apple didn’t respond to a request for comment. 

“If Apple failed with generative AI big time, your average employer will fail too,” AI expert and author Dr. Andriy Burkov wrote. “Work with AI starts with … admitting that it can’t do even 10% of what everyone says it can do.”

I think this, if true, says less about Apple and more about generative AI.

Apple has made a name for itself by perfecting — not inventing — technology. It may be that this is a technology that cannot be perfected. 

As a firm that highly values consumer trust — with its massive install base of nearly 2.4 billion active devices — Apple cannot compete with companies that don’t care about violating user trust, and that don’t care about injecting unreliable, potentially unsafe technology into their user base. Apple is not a social media company. It’s a different breed, playing a different game. 

Apple has much more at stake than any of these other names if an AI integration goes poorly. 

And the others have much more skin in the game, here. Apple isn’t losing anything by slow-playing generative AI; they haven’t poured the billions into it that the rest have, so they don’t need it to generate billions in return. 

They just need to keep selling good, robust, reliable consumer products. 

They don’t need generative AI. 

When Google launched AI Overviews, it told people to put glue on pizza and eat rocks. That’s since stopped, but its regular output of incorrect information has not. When xAI launched Grok 3, users were quickly able to jailbreak the system into outputting hundreds of pages of directions on how to make and sell drugs including lethal nerve agents. Last year, Microsoft’s AI enabled those explicit images of Taylor Swift that went viral online.

And all of these companies are locked in a money-losing endeavor, where the cost of training and operating these models is vastly more significant than any drips and drops of revenue they may be deriving from it. 

All in, this is a situation that Apple can’t, it seems, tolerate. 

I don’t think that’s a bad thing at all; Apple is valued at $3+ trillion not for its hypothetical AI-related potential (like the valuations of Microsoft, OpenAI, Tesla, and Google), but for its incredible ability to sell the world on an ecosystem of interconnected products. 

When Apple does generative AI, I imagine they will do it right — and if it can’t be done right, they won’t do it. 

The AI bubble is barrelling toward a burst. 

When it comes, Apple will not be in the blast zone.

🤔 Your thought process:

Selected Image 1 (Left):

• “The hanging plants look more natural in the real one. Where you can see the anchors.”

Selected Image 2 (Right):

• “The tree in Image 1 seemed too much for real life.”

💭 A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Here’s your view on investing in OpenAI:

Nearly 40% of you think OpenAI is like Theranos; it’ll go down soon, and investing in the startup is a mistake.

23% would consider investing in OpenAI, if the valuation was lower, 15% can’t imagine investing at this valuation and 17% would absolutely invest in the startup.

If you want to get in front of an audience of 450,000+ developers, business leaders and tech enthusiasts, get in touch with us here.