BTC $80,465.34 +0.40%
ETH $2,341.39 +0.84%
BNB $748.62 +0.64%
XRP $1.32 +1.59%
SOL $100.77 -0.16%
TRX $0.3287 -0.45%
DOGE $0.0781 +0.90%
ADA $0.2047 -0.61%
BCH $289.77 +3.71%
LINK $11.45 +0.11%
HYPE $81.20 +2.43%
AAVE $131.88 -2.67%
SUI $0.9218 +0.91%
XLM $0.1694 +0.42%
ZEC $1,293.14 +4.72%
AAPL $346.06 +0.85%
AMZN $254.30 +2.12%
GOOGL $322.14 +1.29%
MSFT $531.11 -0.30%
META $623.35 -1.60%
NVDA $217.75 +0.63%
TSLA $406.34 +3.05%
SNDK $1,298.92 +1.77%
INTC $93.50 +1.80%
SPCX $149.91 +0.21%
MU $933.37 +2.08%
AMD $560.38 +5.19%
BTC $80,465.34 +0.40%
ETH $2,341.39 +0.84%
BNB $748.62 +0.64%
XRP $1.32 +1.59%
SOL $100.77 -0.16%
TRX $0.3287 -0.45%
DOGE $0.0781 +0.90%
ADA $0.2047 -0.61%
BCH $289.77 +3.71%
LINK $11.45 +0.11%
HYPE $81.20 +2.43%
AAVE $131.88 -2.67%
SUI $0.9218 +0.91%
XLM $0.1694 +0.42%
ZEC $1,293.14 +4.72%
AAPL $346.06 +0.85%
AMZN $254.30 +2.12%
GOOGL $322.14 +1.29%
MSFT $531.11 -0.30%
META $623.35 -1.60%
NVDA $217.75 +0.63%
TSLA $406.34 +3.05%
SNDK $1,298.92 +1.77%
INTC $93.50 +1.80%
SPCX $149.91 +0.21%
MU $933.37 +2.08%
AMD $560.38 +5.19%

poll

All
Article
Flash

Apollo Chief Economist: The profit margins of the seven tech giants are rising, while the rest of the S&P companies are stagnant

The chief economist of the American asset management firm Apollo, Torsten Sløk, released five charts, believing that AI has not yet brought significant financial returns to large American companies outside of the tech giants. The per capita revenue and profit margins of the seven tech giants continue to rise; excluding them, the 493 companies in the S&P have seen these two metrics remain basically stagnant, and the data on total factor productivity in the U.S. has not shown significant signs of AI-driven growth.Apollo cited data from the corporate payment platform Ramp, stating that among its clients, the top 10% of companies by spending contributed 99.5% of the model hosting and inference costs. Among the top 1% of companies with the highest AI spending, about 80% have venture capital backgrounds; for companies at the median spending level, this ratio is only 13%. A large amount of AI computing power consumption is concentrated in a few companies, many of which rely on venture capital support.Previous statistics from Apollo indicated that during the second quarter earnings season of 2026, 69% of S&P 500 companies had disclosed that they were actually using AI, but only 2% reported metrics for ongoing tracking of effectiveness. Another NBER study updated in October this year surveyed nearly 9,000 corporate executives from seven countries, with respondents estimating that over the past three years, AI has only improved the annual growth rate of labor productivity in companies by about 0.2 percentage points.

AI Agent Security Risk Exposure: Attackers Can Exploit "Memory Pollution" to Induce Misoperation of Funds

The GoPlus Security team has disclosed a new type of attack in its AgentGuard AI project: inducing AI agents to perform unauthorized sensitive operations through "memory poisoning." This attack method does not rely on traditional vulnerabilities or malicious code but exploits the long-term memory mechanism of AI agents. For example, an attacker first induces the agent to "remember preferences," such as "usually prioritizing proactive refunds instead of waiting for chargebacks," and then uses vague expressions like "process as usual" or "execute as before" in subsequent instructions, thereby triggering automated financial operations.GoPlus points out that the key risk in such cases lies in the AI agent mistakenly treating "historical preferences" as a basis for authorization, leading to financial losses or security incidents in operations such as refunds, transfers, and configuration changes. To address this issue, the team has proposed several protective recommendations, including:Operations involving refunds, transfers, deletions, or sensitive configurations must require explicit confirmation in the current session.Memory-related instructions like "habit," "usual way," and "as before" should be regarded as high-risk state changes.Long-term memory must have a traceability mechanism (writer, time, confirmation status).Vague instructions should automatically elevate the risk level and trigger secondary verification.Long-term memory must not replace real-time authorization processes.The team emphasizes that the "AI agent memory system" should be viewed as a potential attack surface and should be constrained and audited through a dedicated security framework.
app_icon
ChainCatcher Building the Web3 world with innovations.