AI yields no profit for most firms amid bias and dead projects

Zero financial payoff despite massive AI spending

An Apollo economist told TechCrunch that AI has delivered zero financial payoff to corporate America outside the tech sector.¹²³⁴⁵ The statement follows years of headline‑grabbing funding rounds and public promises. The economist’s comment reflects a gap between hype and measurable returns.

TechCrunch’s reporting notes that the lack of profit comes despite “massive investments” and widespread hype. The economist did not provide a dollar figure, but the implication is clear: most non‑tech firms have not turned AI research into revenue. The finding challenges the narrative that AI adoption automatically boosts the bottom line.

Political bias surfaces in leading models

A new TechCrunch study compared the political leanings of two high‑profile language models.⁶⁷⁸ Grok’s AI overwhelmingly recommended Republican candidates, while OpenAI’s GPT favored Democrats.⁶⁷⁸ The two models disagreed in more than 90 % of nearly 500 election races the study examined.⁶⁷⁸

The study’s methodology involved prompting each model with the same set of election‑related queries.⁶⁸ The stark partisan split raises concerns about algorithmic influence on public opinion.⁶⁷⁸ Neither model’s developers offered a detailed explanation for the bias, and the study did not attribute the divergence to a single factor.⁶⁸

Edge AI gains a foothold, but practical value remains limited

Block announced a new voice model that can run locally on a MacBook M4 Max, according to TechCrunch. Running the model on‑device eliminates the need for cloud inference, which could reduce latency and privacy concerns for end users.

The announcement arrived alongside TechCrunch Mobility’s reminder that AI is increasingly embedded in transportation systems. While the voice model showcases progress in edge computing, it does not yet translate into a clear revenue stream for Block. The capability is a technical milestone, not a proven profit driver.

The AI graveyard expands with high‑profile failures

TechCrunch compiled a running list of AI projects that never reached their promised outcomes. The list includes Apple’s repeatedly delayed Siri AI and OpenAI’s messy “super app” launch. Both initiatives generated buzz but failed to deliver functional products on schedule.

These failures illustrate a broader pattern: ambitious AI roadmaps often outpace engineering realities. The “AI graveyard” entry underscores that even well‑funded, brand‑heavy projects can stall. The pattern suggests that hype can obscure technical risk.

Why profit remains elusive and what to watch next

The zero‑payoff finding, the bias study, and the growing catalog of dead projects point to a common theme: AI’s commercial impact is still uncertain. Companies pour capital into research, yet measurable returns lag behind expectations. Technical advances such as on‑device inference show promise but have not yet cracked the profitability barrier.

Regulators and investors will likely focus on three signals in the coming months. First, earnings reports from non‑tech firms that claim AI‑driven growth will be scrutinized for concrete metrics. Second, further bias audits of language models are expected, especially as political stakes rise. Third, the performance of edge AI products like Block’s voice model will be tracked for adoption rates and cost savings.

What to watch: the next quarterly filings from major corporations that tout AI initiatives, the release of any follow‑up bias studies from independent labs, and the market response to Block’s on‑device voice offering. These data points will indicate whether AI moves from a hype‑driven sprint to a sustainable, profit‑generating marathon.

By Maya Chen